This reminds me of a bug report I made about a year and a half ago,
which has seemingly been ignored in all that time ("net-p2p/mldonkey:
duplicate index entry when GUI option is unset
<https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=278934>"). I'm not
sure that it's the same thing (I don't really know enough about the
details of the build system) and it's not exactly the biggest problem in
the world, but I'm wondering if someone would be willing to please take
a look. I suspect that I might have figured out the cause (and I
described my hypothesis in that bug report), but again I don't really
know enough about the build system.
On 12/18/2025 10:26 AM, Mark Millard wrote:
On Dec 18, 2025, at 07:10, void<[email protected]> wrote:
poudriere complains like this after it processes the ports tree:
pkg-static: caldera: duplicate dependency listing: base64
pkg-static: nagstamon: duplicate dependency listing: py311-requests-kerberos
pkg-static: py311-databricks-sql-connector: duplicate dependency listing:
py311-requests-kerberos
pkg-static: py311-hdfs: duplicate dependency listing: py311-requests-kerberos
pkg-static: py311-librt: duplicate dependency listing: base64
pkg-static: py311-ticketutil: duplicate dependency listing:
py311-requests-kerberos
ports is at 729666
For ports main branch:
Sun, 07 Dec 2025:
• git: b3058e362102 - main - devel/base64: Add base64 0.5.2 Po-Chuan Hsieh
. . .
• git: 077d9ba380e5 - main - www/py-requests-kerberos: Add
py-requests-kerberos 0.15.0 Po-Chuan Hsieh
devel/base64 created a second port with the same PKGNAME as long
used by converters/base64 . (The content was not a duplicate.)
www/py-requests-kerberos created a duplicate of
security/py-requests-kerberos (including the PKGNAME).
The fixes are recent:
Wed, 17 Dec 2025
• git: d773ece5bc68 - main - www/py-requests-kerberos: Remove duplicate
port Po-Chuan Hsieh
• git: f554dcbcf289 - main - devel/base64: Change PKGNAME to avoid conflict
with converters/base64 Po-Chuan Hsieh
So, for around 10 days a couple of PKGNAME's were not unique in
the ports tree: the ports tree itself was the problem.
Note:
729666 does not seem to be a commit hash prefix from the main
branch. It is not handy to figure out what commit corresponds
to a "--first-parent --count for merge-base" number, if that
is what it is. dev-commits-ports-main and the likes of
https://cgit.freebsd.org/ports/log/ are not set up for lookups
based on such. If it is such a number, an "n" prefix might be
appropriate.
===
Mark Millard
marklmi at yahoo.com
